Announcements

• Field trip booked for Friday
• Vans leaving Jefferson at 08:30 (Worcester Limo X 2)
• Meet you at 150 Town Farm Road
• Weather depending

Overview of EO Limits

• Physical
• Methodological
• Societal

Physical Limits 1

• Space-time tradeoff
• Initiated by conflict between resolution and extent

Physical Limits 2

• Temporal: length of record

Physical limits 3

• Surface characteristics: terrain and vegetation
• Atmosphere
• Perspective

Methodological Limits

• Inadequate models:
  • Need these to solve the inverse problem
  • Models outpaced by EO data
• Not enough cal/val data
• Competing standards

Societal Limits

• Expertise barrier
• Institutional limitations
  • Data access and collection policies
  • Coordination
• Accessibility to methods
• Collection biases

The Combined Effect of Limits

Innovation Pushing Back Limits 1

• More Eyes in the Sky

Left: max. resolution versus year of launch (Belward & Skøien, 2015), Right: number of launches (Radtke et al, 2017)

Limits Addressed

• Physical:
  • Space-time tradeoff mitigated:
    • Resolution versus extent
    • Atmospheric interference
  • Terrain effects (more radar)
  • Perspective (more radar and lidar)
• Societal:
  • Accessibility (cheaper)

How Our Methods Fit In

• PlanetScope

Innovation Pushing Back Limits 2

• Upping the N

geowiki.org

Limits Addressed

• Physical:
  • Perspective
• Methodological:
  • Inadequate calibration/validation data
  • Better models
• Societal:
  • Accesibility

How Our Methods Fit In

• Crowdsourcing platform
• Arable

Innovation Pushing Back Limits 3

• Peering into the gaps

Gaveau et al (2014)

Limits Addressed

• Physical
  • Space-time tradeoff
  • Perspective
• Methodological
  • Inadequate cal/val
• Societal
  • Cost/accessibility

How Our Methods Fit In

• UAS

Innovations Pushing Back Limits 4

• Advances in sensors/communications

Limits Addressed

• Physical:
  • Space-time tradeoff
  • Perspective
• Methodological
  • Better models
  • Inadequate cal/val
• Societal
  • Cost/accessibility

How Our Methods Fit In

• Sequioa

Innovations Pushing Back Limits 5

• Data and model-data fusion

Luo et al (2018): STAIR

SCYM

Scalable Yield Mapper Results

Limits Addressed

• Physical:
  • Space-time tradeoff
  • Duration
• Methodological
  • Better models

How Our Methods Fit In

• Would be great to implement STAIR!
• Arable Marks + Ebee + Planet + DSSAT crop model & R package

Innovations Pushing Back Limits 6

• Computer vision, machine, and deep learning

Azavea Deep Learning for Semantic Segmentation of Aerial Imagery

Limits Addressed

• Methodological:
  • Inadequate models

How Our Methods Fit In

Lei and Luisa’s slide of Debat et al (2016) method

How Our Methods Fit In

Active learning

Innovations Pushing Back Limits 7

• Cloud-based data and analytical platforms

{width = 80%}

Limits Addressed

• Methodological:
  • Inadequate models
• Societal:
  • Accessibility

How Our Methods Fit In

• Use it to fit STAIR/STARFM
• Process Planet
• Host and view UAS imagery

Installations

• eMotion3
• PIX4D